Remove nth occurrence of aggregated item from an aggregation.

[elrayle]

Items can repeat in aggregation. The syntax for the delete method needs to support removing the nth occurrence.

Example without repeat...

class DummyObject < ActiveFedora::Base
  aggregates :members, predicate: RDFVocabularies::PCDMTerms.hasMember, class_name: "ActiveFedora::Base"
end

do_parent = DummyObject.new
do_child1 = DummyObject.new
do_child2 = DummyObject.new
do_child3 = DummyObject.new

do_parent.members = [do_child1,do_child2,do_child3]   # => [do_child1,do_child2,do_child3]
do_parent.members.delete do_child1   # => do_child1

Example with repeat...

class DummyObject < ActiveFedora::Base
  aggregates :members, predicate: RDFVocabularies::PCDMTerms.hasMember, class_name: "ActiveFedora::Base"
end

do_parent = DummyObject.new
do_child1 = DummyObject.new
do_child2 = DummyObject.new
do_child3 = DummyObject.new

do_parent.members = [do_child1,do_child2,do_child1,do_child3]   # => [do_child1,do_child2,do_child1,do_child3]

# delete 2nd occurrence of do_child1
do_parent.members.delete( do_child1, 2 )   # => do_child1
do_parents.members          # => [do_child1,do_child2,do_child3]

See also...

• PCDM related issue Remove nth occurrence of an object from a collection/object. samvera/hydra-pcdm#102
• PCDM RemoveObjectFromCollection.rb
• collection/remove_object_spec.rb context 'when object repeats' around line 71.

[elrayle]

If nth occurrence is specified and is positive, then count from the beginning of the list.
If nth occurrence is specified and is negative, then count from the end of the list.

[elrayle]

See usage with filtered associations... #45

[jcoyne]

It would more closely match the ruby enumerable API, and be more generally useful if we supported delete_at(n) rather than delete( child, n )

[elrayle]

I am fine with delete_at(n). What would you expect to happen if delete(child) is used and child is in the list multiple times.

• delete all instances of child
• delete first instance of child

Examples...

col1.members = [col1,obj1,obj2,col2,obj1,col1]
deleted_objs = col1.members.delete_at 3   # => [obj2]
col1.members                              # => [col1,obj1,col2,obj1,col1]

col1.members = [col1,obj1,obj2,col2,obj1,col1]
deleted_objs = col1.members.delete obj1   # => [obj1]
col1.members                              # => [col1,obj2,col2,obj1,col1]   OR   
                                          # => [col1,obj2,col2,col1]

col1.members = [col1,obj1,obj2,col2,obj1,col1]
col1.child_objects                              # => [obj1,obj2,obj1]
deleted_objs = col1.child_objects.delete_at 3   # => [obj1]
col1.child_objects                              # => [obj1,obj2]
col1.members                                    # => [col1,obj1,obj2,col2,col1]

col1.members = [col1,obj1,obj2,col2,obj1,col1]
col1.child_objects                              # => [obj1,obj2,obj1]
deleted_objs = col1.child_objects.delete obj1   # => [obj1]
col1.child_objects                              # => [obj2,obj1]   OR   [obj2]

[escowles]

The standard ruby Array behavior is to delete all of them, and use delete_at if you want to delete just one.